Received-SPF: pass (google.com: domain of linux-kernel+bounces-157843-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <b49224dbc89bd304a5c20eef4430dfea40f4b6a5.camel@gmail.com>
Subject: Re: [RFC PATCH bpf-next] libbpf: print character arrays as strings
 if possible
From: Eduard Zingerman <eddyz87@gmail.com>
To: Quentin Deslandes <qde@naccy.de>, bpf@vger.kernel.org
Cc: Andrii Nakryiko <andrii@kernel.org>, Alexei Starovoitov
 <ast@kernel.org>,  Daniel Borkmann <daniel@iogearbox.net>, Martin KaFai Lau
 <martin.lau@linux.dev>, Song Liu <song@kernel.org>,  Yonghong Song
 <yonghong.song@linux.dev>, John Fastabend <john.fastabend@gmail.com>, KP
 Singh <kpsingh@kernel.org>, Stanislav Fomichev <sdf@google.com>, Hao Luo
 <haoluo@google.com>, Jiri Olsa <jolsa@kernel.org>,
 linux-kernel@vger.kernel.org,  kernel-team@meta.com
Date: Wed, 24 Apr 2024 16:33:27 -0700
In-Reply-To: <20240413213904.146261-1-qde@naccy.de>
References: <20240413213904.146261-1-qde@naccy.de>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.4-0ubuntu2 
Precedence: bulk
MIME-Version: 1.0

On Sat, 2024-04-13 at 23:39 +0200, Quentin Deslandes wrote:
> Introduce the new print_strings flag in btf_dump_type_data_opts. If
> enabled, libbpf will print character arrays as strings if they meet the
> following conditions:
> - Contains a nul-termination character ('\0')
> - Contains only printable characters before the nul-termination character
>=20
> If print_strings is set to false (default value), the existing
> behavior remains unchanged.
>=20
> With print_strings=3Dfalse:
> .str_array =3D (__u8[14])[
>     'H',
>     'e',
>     'l',
>     'l',
>     'o',
> ],
>=20
> With print_strings=3Dtrue:
> .str_array =3D (__u8[14])"Hello",
>=20
> Signed-off-by: Quentin Deslandes <qde@naccy.de>
> ---

Hi Quentin,

Thank you for this patch, sorry for the delay reviewing it.
Could you please also add a few tests in
tools/testing/selftests/bpf/prog_tests/btf_dump.c ?

[...]

> @@ -2021,6 +2022,21 @@ static int btf_dump_var_data(struct btf_dump *d,
>  	return btf_dump_dump_type_data(d, NULL, t, type_id, data, 0, 0);
>  }
>=20
> +static bool btf_dump_isprint_str(const char *data, unsigned int len)
> +{
> +	unsigned int i;
> +
> +	for (i =3D 0; i < len; ++i) {
> +		if (data[i] =3D=3D '\0')
> +			return true;
> +
> +		if (!isprint(data[i]))
> +			return false;

Would it make sense to use isprint_l() and specify something like C locale?=
=20

> +	}
> +
> +	return false;
> +}
> +
>  static int btf_dump_array_data(struct btf_dump *d,
>  			       const struct btf_type *t,
>  			       __u32 id,
> @@ -2047,8 +2063,14 @@ static int btf_dump_array_data(struct btf_dump *d,
>  		 * char arrays, so if size is 1 and element is
>  		 * printable as a char, we'll do that.
>  		 */
> -		if (elem_size =3D=3D 1)
> +		if (elem_size =3D=3D 1) {
>  			d->typed_dump->is_array_char =3D true;
> +			if (d->typed_dump->print_strings &&
> +					btf_dump_isprint_str(data, array->nelems)) {
> +				btf_dump_type_values(d, "\"%s\"", data);

Note: this would have to deal with escape sequences,
otherwise strings containing '\' would be printed incorrectly.

> +				return 0;
> +			}
> +		}
>  	}
>=20
>  	/* note that we increment depth before calling btf_dump_print() below;

[...]